CN105207272A - Electric power system dynamic random economic dispatching method and device based on general distribution - Google Patents

Abstract

The invention discloses an electric power system dynamic random economic dispatching method and device based on general distribution. The method comprises the steps that day-ahead system load prediction data, day-ahead wind power prediction data and the like are input, it is assumed that a planned output of a wind power plant is a wind power prediction value, the quadratic programming problem is solved to obtain all thermal power unit outputs based on wind power prediction, the wind power prediction value and all the solved thermal power unit outputs are used as initial iteration points of an interior point method, the interior point method is utilized for iteratively solving a converted convex optimization problem with a constraint condition being linearity until iteration is stopped, and the planned outputs of the thermal power units and the planned output of the wind power plant are output. The method has high popularization value and good application prospects.

Description

The random economic dispatch method of Electrical Power System Dynamic based on general distribution and device

Technical field

The invention belongs to operation and control of electric power system field, relate to the random economic dispatch technical scheme of a kind of consideration wind-powered electricity generation based on general distribution Electrical Power System Dynamic that is low, that over-evaluate cost.

Background technology

Along with the large-scale grid connection of wind-powered electricity generation, the uncertainty of wind power brings new challenge to the economic dispatch of electric power system.Along with the increase gradually of wind-powered electricity generation permeability, how rationally to describe the uncertainty of wind power and applied in the economic dispatch of electric power system and optimizing operation significant.

Random optimization is the effective ways of a kind of process containing uncertainty optimization problem, has been widely used at present containing in probabilistic Economic Dispatch problem.How accurate description wind power uncertainty, effectively to solve corresponding Optimized model be key issue containing the random economic dispatch of wind-powered electricity generation electric power system.

Analyze Economic Dispatch based on randomized optimization process, and then obtain considering the electric power system fired power generating unit of wind-powered electricity generation predicated error and the plan of exerting oneself of wind energy turbine set, Chinese scholars has carried out large quantity research to this, research method is broadly divided into two classes:

(1) based on the probabilistic random economic dispatch method of wind speed, the method is based on the historical data of wind farm wind velocity, by portraying wind speed is probabilistic, utilize wind speed-wind power curve to obtain the distribution of wind power, and then set up corresponding random economic dispatch method.Usually these class methods portray the distribution of wind speed relatively accurately, but are carried out the distribution of approximate description wind power by the piecewise function of power characteristic, can increase the error of fitting of wind power distribution, the accuracy of the corresponding random economic dispatch model of impact.

(2) based on the probabilistic random economic dispatch method of wind power, the method is based on the historical data of wind energy turbine set wind power, directly portray the uncertainty of wind power, obtain the distributed constant of actual wind power, and then set up corresponding random economic dispatch method.Usually these class methods can be avoided transforming by wind speed-wind power the error brought, but do not have most suitable distribution function to describe the distribution of wind power, and the solution procedure more complicated of the random optimization of correspondence.

Generally speaking, Weibull distributed model is commonly used in the distribution describing wind speed, and normal distribution is commonly used in the distribution describing wind power.The accuracy described wind power distribution is closely related with the accuracy of corresponding stochastic and dynamic economic dispatch solution to model.But not yet there is the relevant technical scheme with practical value at present.

Summary of the invention

The present invention is directed to the defect of prior art, the random economic dispatch technical scheme of Electrical Power System Dynamic based on general distribution is provided.

Technical solution of the present invention provides a kind of Electrical Power System Dynamic based on general distribution random economic dispatch method, comprises the following steps:

Step 1, input system loading prediction data and a few days ago wind power prediction data, thermal power unit operation parameter, system line parameter, historical statistical data a few days ago, described historical statistical data comprises the general profile parameter of actual wind power under different wind power prediction level, beta, gamma;

Step 2, if p _i,tbe exerting oneself of i-th fired power generating unit t, fired power generating unit add up to I, i=1,2 ..., I, w _j,tfor the plan of a jth wind energy turbine set t is exerted oneself, wind energy turbine set add up to J, j=1,2 ..., J, T are the sum in moment, t=1,2 ..., T,

The plan of false wind electric field is exerted oneself w _j,tfor wind power prediction value w _{j, fcst, t}, wind power prediction value w _{j, fcst, t}thered is provided by wind power prediction data a few days ago, solve following quadratic programming problem with Novel Algorithm, obtain exerting oneself p based on each fired power generating unit of prediction wind power _i,t ⁽⁰⁾,

$min\underset{t=1}{\overset{T}{\Σ}}\underset{i=1}{\overset{I}{\Σ}}(a_i{p}_{i,t}^2+b_i{p}_{i,t}+c_i)+C_{wind}$ (formula one)

Wherein, C _windfor the total cost of wind-powered electricity generation, a _i, b _i, c _iit is the fuel cost coefficient of i-th fired power generating unit;

$\underset{i=1}{\overset{I}{\Σ}}{p}_{i,t}+\underset{j=1}{\overset{J}{\Σ}}{w}_{j,t}=L_t,\∀t$ (formula two)

${\mathrm{\η}}_{i,t}{p}_{min,i}\≤p_{i,t}\≤{\mathrm{\η}}_{i,t}{p}_{max,i},\∀i,t$ (formula three)

$0\≤w_{j,t}\≤w_{r,j},\∀j,t$ (formula four)

$p_{i,t}-p_{i,t-1}\≤{\mathrm{\η}}_{i,t-1}\·r_{u,max,i}+({\mathrm{\η}}_{i,t}-{\mathrm{\η}}_{i,t-1})p_{min,i}+(1-{\mathrm{\η}}_{i,t})p_{max,i},\∀i,t$ (formula five)

$p_{i,t-1}-p_{i,t}\≤{\mathrm{\η}}_{i,t}\·r_{d,max,i}+({\mathrm{\η}}_{i,t-1}-{\mathrm{\η}}_{i,t})p_{min,i}+(1-{\mathrm{\η}}_{i,t-1})p_{max,i},\∀i,t$ (formula six)

$0\≤r_{u,i,t}\≤min\{{\mathrm{\η}}_{i,t}{p}_{max,i}-p_{i,t},{\mathrm{\η}}_{i,t}{r}_{u,max,i}\},\∀i,t$ (formula seven)

$0\≤r_{d,i,t}\≤min\{p_{i,t}-{\mathrm{\η}}_{i,t}{p}_{min,i},{\mathrm{\η}}_{i,t}{r}_{d,max,i}\},\∀i,t$ (formula eight)

Wherein, L _tfor the total load of t system, provided by system loading prediction data a few days ago; η _i,tbe the on-off state of i-th fired power generating unit t, r _{u, max, i}and r _{d, max, i}be respectively i-th fired power generating unit maximum creep speed up and down, p _{min, i}and p _{max, i}be minimum load and the maximum output of i-th fired power generating unit, provided by thermal power unit operation parameter; w _r,jfor the installed capacity of wind-driven power of a jth wind energy turbine set, r _{u, i, t}and r _{d, i, t}it is the reserve capacity up and down of i-th fired power generating unit t;

$-(1-\mathrm{\μ})\·F^{max}\≤F_t\≤(1-\mathrm{\μ})\·F^{max},\∀t$ (formula nine)

Wherein, F _tfor the vector of each Line Flow of t; F ^maxfor the vector of each circuit maximum transfer capacity, the reserved transmission capacity of μ to be transmission line be wind-powered electricity generation fluctuation accounts for the ratio of each branch road maximum transfer capacity, is provided by system line parameter;

$\underset{j=1}{\overset{J}{\Σ}}{w}_{j,t}-\underset{i=1}{\overset{I}{\Σ}}{r}_{u,i,t}\≤F_{\mathrm{\Σ},t}^{-1}(1-c_u),\∀t$ (formula ten)

$-\underset{j=1}{\overset{J}{\Σ}}{w}_{j,t}-\underset{i=1}{\overset{I}{\Σ}}{r}_{d,i,t}\≤-F_{\mathrm{\Σ},t}^{-1}\left(c_d\right),\∀t$ (formula 11)

Wherein, for the inverse function of the CDF that all wind energy turbine set actual capabilities of t are exerted oneself, c _uand c _dbe respectively the confidence level that corresponding constraints meets; Described CDF is the cumulative distribution function of general distribution, and according to the corresponding general profile parameter of t, beta, gamma is determined;

Step 3, by wind power prediction value w _{j, fcst, t}to exert oneself p with each fired power generating unit solved _i,t ⁽⁰⁾as the primary iteration point x of interior point method ⁽⁰⁾;

Step 4, arranges convergence criterion parameter ε and the maximum iteration time N of interior point method _iter;

Step 5, the constraints after utilizing interior point method iterative to transform is linear convex optimization problem, and namely by formula 12, the convex optimization problem that formula two-Shi 11 is formed, until meet convergence criterion parameter ε or maximum iteration time N _itertime iteration stopping, enter step 6,

${\mathrm{minC}}_{all}(p_{i,t},w_{j,t})=\underset{t=1}{\overset{T}{\Σ}}\underset{i=1}{\overset{I}{\Σ}}{C}_{g,i,t}\left(p_{i,t}\right)+\underset{t=1}{\overset{T}{\Σ}}\underset{j=1}{\overset{J}{\Σ}}\left(C_{w,j,t}\right(w_{j,t})+C_{un,j,t}(w_{j,t})+C_{ov,j,t}(w_{j,t}\left)\right)$ (formula 12)

Wherein, C _allfor the operating cost that system is total, C _{g, i, t}for the fuel cost of t i-th fired power generating unit, C _{w, j, t}for the operating cost of a t jth wind energy turbine set, C _{un, j, t}on average cost is underestimated, C for a t jth wind energy turbine set wind power prediction _{ov, j, t}on average cost is over-evaluated for a t jth wind energy turbine set wind power prediction;

Step 6, according to the iteration result of step 5, the plan exporting fired power generating unit and wind energy turbine set is exerted oneself.

And, described fuel cost C _{g, i, t}calculate in the following ways,

$C_{g,i,t}\left(p_{i,t}\right)=a_i{p}_{i,t}^2+b_i{p}_{i,t}+c_i$ (formula 13)

Wherein, p _i,tbe exerting oneself of i-th fired power generating unit t, a _i, b _i, c _iit is the fuel cost coefficient of i-th fired power generating unit.

And, described operating cost C _{w, j, t}calculate in the following ways,

C _{w, j, t}(w _j,t)=d _jw _j,t(formula 14)

Wherein, w _j,tfor the plan of a jth wind energy turbine set t is exerted oneself, d _jfor the operating cost coefficient of a jth wind energy turbine set.

And, describedly on average underestimate cost C _{un, j, t}what adopt wind energy turbine set on average abandons eolian, calculates in the following ways,

$C_{un,j,t}\left(w_{j,t}\right)=k_{un,j}\underset{w_{j,t}}{\overset{w_{r,j}}{\∫}}(w_{av,j,t}-w_{j,t})\·f_j\left(w_{av,j,t}\right){\mathrm{dw}}_{av,j,t}$ (formula 15)

Wherein, k _{un, j}for a jth wind energy turbine set corresponding underestimate cost coefficient, w _{av, j, t}for the actual capabilities of a jth wind energy turbine set t are exerted oneself, f _j(w _{av, j, t}) be the jth wind energy turbine set probability density function that actual capabilities are exerted oneself under corresponding wind-powered electricity generation prediction level, expression-form is the probability density function of general distribution, and according to the corresponding general profile parameter of t, beta, gamma is determined.

And, describedly on average over-evaluate cost C _{ov, j, t}the average stand-by cost of employing system, calculates in the following ways,

$C_{ov,j,t}\left(w_{j,t}\right)=k_{ov,j}\underset{0}{\overset{w_{j,t}}{\∫}}(w_{j,t}-w_{av,j,t})\·f_j\left(w_{av,j,t}\right){\mathrm{dw}}_{av,j,t}$ (formula 16)

Wherein, k _{ov, j}for a jth wind energy turbine set corresponding over-evaluate cost coefficient, w _{av, j, t}for the actual capabilities of a jth wind energy turbine set t are exerted oneself, f _j(w _{av, j, t}) be the jth wind energy turbine set probability density function that actual capabilities are exerted oneself under corresponding wind-powered electricity generation prediction level, expression-form is the probability density function of general distribution, and according to the corresponding general profile parameter of t, beta, gamma is determined.

The present invention is corresponding provides a kind of Electrical Power System Dynamic based on general distribution random economic dispatch system, comprises with lower module:

Input module, for inputting system loading prediction data and a few days ago wind power prediction data, thermal power unit operation parameter, system line parameter, historical statistical data a few days ago, described historical statistical data comprises the general profile parameter of actual wind power under different wind power prediction level, beta, gamma;

Tentatively solve module, for establishing p _i,tbe exerting oneself of i-th fired power generating unit t, fired power generating unit add up to I, i=1,2 ..., I, w _j,tfor the plan of a jth wind energy turbine set t is exerted oneself, wind energy turbine set add up to J, j=1,2 ..., J, T are the sum in moment, t=1,2 ..., T,

The plan of false wind electric field is exerted oneself w _j,tfor wind power prediction value w _{j, fcst, t}, wind power prediction value w _{j, fcst, t}thered is provided by wind power prediction data a few days ago, solve following quadratic programming problem with Novel Algorithm, obtain exerting oneself p based on each fired power generating unit of prediction wind power _i,t ⁽⁰⁾,

$min\underset{t=1}{\overset{T}{\Σ}}\underset{i=1}{\overset{I}{\Σ}}(a_i{p}_{i,t}^2+b_i{p}_{i,t}+c_i)+C_{wind}$ (formula one)

Wherein, C _windfor the total cost of wind-powered electricity generation, a _i, b _i, c _iit is the fuel cost coefficient of i-th fired power generating unit;

$\underset{i=1}{\overset{I}{\Σ}}{p}_{i,t}+\underset{j=1}{\overset{J}{\Σ}}{w}_{j,t}=L_t,\∀t$ (formula two)

${\mathrm{\η}}_{i,t}{p}_{min,i}\≤p_{i,t}\≤{\mathrm{\η}}_{i,t}{p}_{max,i},\∀i,t$ (formula three)

$0\≤w_{j,t}\≤w_{r,j},\∀j,t$ (formula four)

$p_{i,t}-p_{i,t-1}\≤{\mathrm{\η}}_{i,t-1}\·r_{u,max,i}+({\mathrm{\η}}_{i,t}-{\mathrm{\η}}_{i,t-1})p_{min,i}+(1-{\mathrm{\η}}_{i,t})p_{max,i},\∀i,t$ (formula five)

$p_{i,t-1}-p_{i,t}\≤{\mathrm{\η}}_{i,t}\·r_{d,max,i}+({\mathrm{\η}}_{i,t-1}-{\mathrm{\η}}_{i,t})p_{min,i}+(1-{\mathrm{\η}}_{i,t-1})p_{max,i},\∀i,t$ (formula six)

$0\≤r_{u,i,t}\≤min\{{\mathrm{\η}}_{i,t}{p}_{max,i}-p_{i,t},{\mathrm{\η}}_{i,t}{r}_{u,max,i}\},\∀i,t$ (formula seven)

$0\≤r_{d,i,t}\≤min\{p_{i,t}-{\mathrm{\η}}_{i,t}{p}_{\min ,i},{\mathrm{\η}}_{i,t}{r}_{d,max,i}\},\∀i,t$ (formula eight)

Wherein, L _tfor the total load of t system, provided by system loading prediction data a few days ago; η _i,tbe the on-off state of i-th fired power generating unit t, r _{u, max, i}and r _{d, max, i}be respectively i-th fired power generating unit maximum creep speed up and down, p _{min, i}and p _{max, i}be minimum load and the maximum output of i-th fired power generating unit, provided by thermal power unit operation parameter; w _r,jfor the installed capacity of wind-driven power of a jth wind energy turbine set, r _{u, i, t}and r _{d, i, t}it is the reserve capacity up and down of i-th fired power generating unit t;

$-(1-\mathrm{\μ})\·F^{max}\≤F_t\≤(1-\mathrm{\μ})\·F^{max},\∀t$ (formula nine)

Wherein, F _tfor the vector of each Line Flow of t; F ^maxfor the vector of each circuit maximum transfer capacity, the reserved transmission capacity of μ to be transmission line be wind-powered electricity generation fluctuation accounts for the ratio of each branch road maximum transfer capacity, is provided by system line parameter;

$\underset{j=1}{\overset{J}{\Σ}}{w}_{j,t}-\underset{i=1}{\overset{I}{\Σ}}{r}_{u,i,t}\≤F_{\mathrm{\Σ},t}^{-1}(1-c_u),\∀t$ (formula ten)

$-\underset{j=1}{\overset{J}{\Σ}}{w}_{j,t}-\underset{i=1}{\overset{I}{\Σ}}{r}_{d,i,t}\≤-F_{\mathrm{\Σ},t}^{-1}\left(c_d\right),\∀t$ (formula 11)

Wherein, for the inverse function of the CDF that all wind energy turbine set actual capabilities of t are exerted oneself, c _uand c _dbe respectively the confidence level that corresponding constraints meets; Described CDF is the cumulative distribution function of general distribution, and according to the corresponding general profile parameter of t, beta, gamma is determined;

Initialization module, for by wind power prediction value w _{j, fcst, t}to exert oneself p with each fired power generating unit solved _i,t ⁽⁰⁾as the primary iteration point x of interior point method ⁽⁰⁾;

Condition setting module, for arranging convergence criterion parameter ε and the maximum iteration time N of interior point method _iter;

Iteration module, the constraints after transforming for utilizing interior point method iterative is linear convex optimization problem, and namely by formula 12, the convex optimization problem that formula two-Shi 11 is formed, until meet convergence criterion parameter ε or maximum iteration time N _itertime iteration stopping, order output module work,

${\mathrm{minC}}_{all}(p_{i,t},w_{j,t})=\underset{t=1}{\overset{T}{\Σ}}\underset{i=1}{\overset{I}{\Σ}}{C}_{g,i,t}\left(p_{i,t}\right)+\underset{t=1}{\overset{T}{\Σ}}\underset{j=1}{\overset{J}{\Σ}}\left(C_{w,j,t}\right(w_{j,t})+C_{un,j,t}(w_{j,t})+C_{ov,j,t}(w_{j,t}\left)\right)$ (formula 12)

Wherein, C _allfor the operating cost that system is total, C _{g, i, t}for the fuel cost of t i-th fired power generating unit, C _{w, j, t}for the operating cost of a t jth wind energy turbine set, C _{un, j, t}on average cost is underestimated, C for a t jth wind energy turbine set wind power prediction _{ov, j, t}on average cost is over-evaluated for a t jth wind energy turbine set wind power prediction;

Output module, for the iteration result according to iteration module, the plan exporting fired power generating unit and wind energy turbine set is exerted oneself.

And, described fuel cost C _{g, i, t}calculate in the following ways,

$C_{g,i,t}\left(p_{i,t}\right)=a_i{p}_{i,t}^2+b_i{p}_{i,t}+c_i$ (formula 13)

Wherein, p _i,tbe exerting oneself of i-th fired power generating unit t, a _i, b _i, c _iit is the fuel cost coefficient of i-th fired power generating unit.

And, described operating cost C _{w, j, t}calculate in the following ways,

C _{w, j, t}(w _j,t)=d _jw _j,t(formula 14)

Wherein, w _j,tfor the plan of a jth wind energy turbine set t is exerted oneself, d _jfor the operating cost coefficient of a jth wind energy turbine set.

And, describedly on average underestimate cost C _{un, j, t}what adopt wind energy turbine set on average abandons eolian, calculates in the following ways,

$C_{un,j,t}\left(w_{j,t}\right)=k_{un,j}\underset{w_{j,t}}{\overset{w_{r,j}}{\∫}}(w_{av,j,t}-w_{j,t})\·f_j\left(w_{av,j,t}\right){\mathrm{dw}}_{av,j,t}$ (formula 15)

Wherein, k _{un, j}for a jth wind energy turbine set corresponding underestimate cost coefficient, w _{av, j, t}for the actual capabilities of a jth wind energy turbine set t are exerted oneself, f _j(w _{av, j, t}) be the jth wind energy turbine set probability density function that actual capabilities are exerted oneself under corresponding wind-powered electricity generation prediction level, expression-form is the probability density function of general distribution, and according to the corresponding general profile parameter of t, beta, gamma is determined.

And, describedly on average over-evaluate cost C _{ov, j, t}the average stand-by cost of employing system, calculates in the following ways,

$C_{ov,j,t}\left(w_{j,t}\right)=k_{ov,j}\underset{0}{\overset{w_{j,t}}{\∫}}(w_{j,t}-w_{av,j,t})\·f_j\left(w_{av,j,t}\right){\mathrm{dw}}_{av,j,t}$ (formula 16)

Wherein, k _{ov, j}for a jth wind energy turbine set corresponding over-evaluate cost coefficient, w _{av, j, t}for the actual capabilities of a jth wind energy turbine set t are exerted oneself, f _j(w _{av, j, t}) be the jth wind energy turbine set probability density function that actual capabilities are exerted oneself under corresponding wind-powered electricity generation prediction level, expression-form is the probability density function of general distribution, and according to the corresponding general profile parameter of t, beta, gamma is determined.

The present invention utilizes general distribution to feature the uncertainty of wind power, based on this, propose the economic dispatch of the dynamic random a few days ago technical scheme considering that wind-powered electricity generation is low, over-evaluate cost, comprise based on the history wind power data of wind energy turbine set, utilize the distribution of actual wind power under the different wind power prediction level of general distributed model matching, take into account the punishment cost that wind-powered electricity generation predicated error is brought, establish that consideration wind-powered electricity generation based on general distribution is low, the Stochastic Optimization Model of the dynamic economic dispatch a few days ago of over-evaluating cost; By transforming and analyzing, the Stochastic Optimization Model of correspondence is converted into a constraints for linear, target function be nonlinear convex optimization problem; And solving corresponding Economic Dispatch Problem in conjunction with Novel Algorithm and interior point method, the plan a few days ago obtaining fired power generating unit and wind energy turbine set is exerted oneself.Empirical tests, the validity of known technical solution of the present invention, has good promotional value and application prospect.

Accompanying drawing explanation

Fig. 1-1 is impact (β=1, the γ=0) figure of general profile parameter on general distribution shape of the embodiment of the present invention.

Fig. 1-2 is impact (α=1, the γ=0) figure of general distributed constant β on general distribution shape of the embodiment of the present invention.

Fig. 1-3 is general distributed constant γ impact (α=1, β=1) figure on general distribution shape of the embodiment of the present invention.

Fig. 2 is the solution procedure figure of the stochastic and dynamic economic dispatch model of the embodiment of the present invention.

Fig. 3 is the IEEE30 node system network topological diagram of the embodiment of the present invention.

Fig. 4 is the switching on and shutting down of the fired power generating unit a few days ago planning chart of the embodiment of the present invention.

Fig. 5-1 be the prediction wind power of the embodiment of the present invention in the 2nd case time actual wind power fitting of distribution design sketch.

Fig. 5-2 be the prediction wind power of the embodiment of the present invention in the 10th case time actual wind power fitting of distribution design sketch.

Fig. 5-3 be the prediction wind power of the embodiment of the present invention in the 19th case time actual wind power fitting of distribution design sketch.

Fig. 6 is the wind power output curve chart of the embodiment of the present invention.

Fig. 7 is the system reserve capacity figure of the randomized optimization process based on normal distribution of the embodiment of the present invention.

Fig. 8 is the system reserve capacity figure of the randomized optimization process based on general distribution of the embodiment of the present invention.

Fig. 9 is each cost curve figure of system under the different confidence levels of the embodiment of the present invention.

Embodiment

In order to make the object of the embodiment of the present invention, technical scheme, advantage more clear, introduce technical scheme of the present invention below in conjunction with the embodiment of the present invention and accompanying drawing.

Technical scheme provided by the invention is a kind of economic dispatch of the dynamic random a few days ago method that consideration wind-powered electricity generation based on general distribution is low, over-evaluate cost, and principle is as follows:

By the prediction of integrated wind plant history wind power and measured data standardization, according to the difference of the predicted value of wind power, branch mailbox is carried out to history wind power data, under different wind power prediction level, survey the distribution of wind power under utilizing the different pre-measuring tank of general distribution function matching, obtain corresponding general distributed constant;

The restriction of exerting oneself of the constraint of consideration system active power balance, fired power generating unit and wind energy turbine set, fired power generating unit ramping rate constraints, system reserve capacity constraint and Line Flow constraint, set up and take into account based on general distribution the economic dispatch of the stochastic and dynamic a few days ago model containing wind-powered electricity generation electric power system that wind-powered electricity generation is low, over-evaluate cost;

Based on the certainty economic dispatch model of prediction wind power, Novel Algorithm is utilized to solve, and the primary iteration point of tried to achieve solution is low as the consideration wind-powered electricity generation solving general distribution, the to over-evaluate cost economic dispatch of stochastic and dynamic a few days ago model;

By transforming and analyzing, stochastic and dynamic Economic Dispatch Problem based on general distribution being changed into constraints is linear convex optimization problem, utilize primary iteration point, solved by interior point method, obtain the optimal solution of the stochastic and dynamic economic dispatch based on general distribution, export the plan power curve of fired power generating unit and wind energy turbine set a few days ago.

First, for the sake of ease of implementation, the Stochastic Optimization Model based on general distribution is introduced:

1.1 general distributed models

The probability density function (ProbabilityDensityFunction, PDF) of general distribution is

$f\left(x\right)=\frac{{\mathrm{\&alpha;\&beta;e}}^{-\mathrm{\&alpha;}(x-\mathrm{\&gamma;})}}{{(1+e^{-\mathrm{\&alpha;}(x-\mathrm{\&gamma;})})}^{\mathrm{\&beta;}+1}},-\mathrm{\&infin;}<x<+\mathrm{\&infin;}---\left(1\right)$

Wherein, x is stochastic variable, and e is math constant, profile parameter, and beta, gamma meets α >0, β >0 ,-∞ < γ <+ ∞.The cumulative distribution function (CumulativeDensityFunction, CDF) of general distribution is

F(x)＝(1+e ^-α(x-γ)) ^-β(2)

The inverse function of its correspondence is

$F^{-1}\left(y\right)=\mathrm{\&gamma;}-\frac{1}{\mathrm{\&alpha;}}ln(y^{-1/\mathrm{\&beta;}}-1)---\left(3\right)$

Wherein, y is cumulative probability.

On the impact of distribution shape as shown in Figure 1, abscissa X is the value of stochastic variable to general distributed constant, and ordinate is probability density.As seen from the figure, α is scale parameter, and α is larger, and scatter less, α is less, scatters larger; β is degree of bias parameter, as 0< β <1, is distributed as left avertence distribution, when β=1, is distributed as without distribution partially, as β >1, is distributed as right avertence distribution; γ is location parameter, and when α and β is constant, different γ only changes the position of general distribution, does not change its shape.

The size of parameter beta affects the offset characteristic of general distribution function.When predict wind power less or larger time, actual wind power output is by the restriction of wind energy turbine set minimum load and maximum output, and its distribution has offset characteristic; When predicting that wind power is placed in the middle, the distribution of actual wind power, close to without distribution partially, namely has symmetry.General distribution can according to the distribution character of actual wind power, by the distribution of actual wind power under the adjustment of the parameter beta well different wind power prediction level of matching.Therefore, when describing the distribution of actual wind power, general distribution has higher using value.

1.2 Stochastic Optimization Model

Based on the Stochastic Optimization Model of chance constraint such as formula shown in (4):

$\left\{\begin{array}{c}\min Ef(x,\mathrm{\&xi;})\\ s.t.\\ \mathrm{Pr}\{g_i(x,\mathrm{\&xi;})\&le;0,i=1,2,...,p)\}\&GreaterEqual;c\\ h_j\left(x\right)\&le;0,j=1,2,...,q\end{array}\right.---\left(4\right)$

Wherein, f (x, ξ) is target function, and x is decision vector, and ξ is random vector, and E is the expectation operator about ξ, and g is the inequality constraints containing random vector, and p is the number of corresponding random constraints, g _i(x, ξ) is i-th inequality constraints containing random vector, and Pr{.} is the probability of corresponding constraint satisfaction, and h is not containing the inequality constraints of random vector, and q is the number of corresponding certainty constraint, h _jx inequality constraints that () does not contain random vector for jth is individual, c is the confidence level of satisfied corresponding inequality constraints.When the distribution function of covariance matrix of sample ξ and the inverse function of CDF thereof have analytical expression, the Stochastic Optimization Model based on chance constraint can be converted into following deterministic models and solve.

$\left\{\begin{array}{c}min\&Integral;f(x,\mathrm{\&xi;})\&CenterDot;p\left(\mathrm{\&xi;}\right)\&CenterDot;d\mathrm{\&xi;}\\ s.t.\\ g_i^{\&prime;}\left(x\right)\&le;F^{-1}\left(c\right),i=1,2,...,p\\ h_j\left(x\right)\&le;0,j=1,2,...,q\end{array}\right.---\left(5\right)$

Wherein, the distribution function that p (ξ) is ξ, g _i' (x) be containing the constraint of stochastic variable after transforming.F ^-1for the inverse function of corresponding CDF.If its constraints is linearly, target function is non-linear, then it is linear nonlinear optimal problem that corresponding optimization problem changes into constraints, can solve with corresponding nonlinear optimization method.

Inverse function due to the CDF of general distribution has closed analytical expression, and thus general distributed model can transform chance constraint effectively, is convenient to solving of corresponding Stochastic Optimization Model.

Then, the economic dispatch model of stochastic and dynamic a few days ago based on general distribution and the method for solving thereof of embodiment of the present invention foundation is introduced:

Stochastic and dynamic economic dispatch model before 2.1 days

Stochastic and dynamic economic dispatch energy safeguards system containing wind-powered electricity generation electric power system meets related constraint under certain confidence level, makes the desired value of system total operating cost minimum.

2.1.2 target function

Consider underestimating and over-evaluate and bringing certain impact to the safety and stability of system of wind-powered electricity generation, the total cost of economic dispatch model herein comprises the punishment cost that the fuel cost of fired power generating unit, the operating cost of wind energy turbine set and wind-powered electricity generation forecasting inaccuracy bring, shown in (6):

${\mathrm{minC}}_{all}(p_{i,t},w_{j,t})=\underset{t=1}{\overset{T}{\&Sigma;}}\underset{i=1}{\overset{I}{\&Sigma;}}{C}_{g,i,t}\left(p_{i,t}\right)+\underset{t=1}{\overset{T}{\&Sigma;}}\underset{j=1}{\overset{J}{\&Sigma;}}\left(C_{w,j,t}\right(w_{j,t})+C_{un,j,t}(w_{j,t})+C_{ov,j,t}(w_{j,t}\left)\right)---\left(6\right)$

Wherein, p _i,tbe exerting oneself of i-th fired power generating unit t, fired power generating unit add up to I, i=1,2 ..., I, w _j,tfor the plan of a jth wind energy turbine set t is exerted oneself, wind energy turbine set add up to J, j=1,2 ..., J, T are the sum in moment, t=1,2 ..., T.C _allfor the operating cost that system is total, C _{g, i, t}for the fuel cost of t i-th fired power generating unit, C _{w, j, t}for the operating cost of a t jth wind energy turbine set, C _{un, j, t}on average underestimate cost for a t jth wind energy turbine set wind power prediction, actual corresponding be wind energy turbine set on average abandon eolian, C _{ov, j, t}on average over-evaluate cost for a t jth wind energy turbine set wind power prediction, actual corresponding is system is that holding power balance enables average stand-by cost for subsequent use.The spinning reserve capacity generally referring to fired power generating unit for subsequent use of electric power system.Namely fired power generating unit existing go out under force level, within a certain period of time can adjustment amount up or down.The expression formula that each cost is corresponding is as follows:

$C_{g,i,t}\left(p_{i,t}\right)=a_i{p}_{i,t}^2+b_i{p}_{i,t}+c_i---\left(7\right)$

C _w,j,t(w _j,t)＝d _jw _j,t(8)

$C_{un,j,t}\left(w_{j,t}\right)=k_{un,j}\underset{w_{j,t}}{\overset{w_{r,j}}{\&Integral;}}(w_{av,j,t}-w_{j,t})\&CenterDot;f_j\left(w_{av,j,t}\right){\mathrm{dw}}_{av,j,t}---\left(9\right)$

$C_{ov,j,t}\left(w_{j,t}\right)=k_{ov,j}\underset{0}{\overset{w_{j,t}}{\&Integral;}}(w_{j,t}-w_{av,j,t})\&CenterDot;f_j\left(w_{av,j,t}\right){\mathrm{dw}}_{av,j,t}---\left(10\right)$

Wherein, a _i, b _i, c _ibe the fuel cost coefficient of i-th fired power generating unit, d _jfor the operating cost coefficient of a jth wind energy turbine set, k _{un, j}, k _{ov, j}for corresponding the underestimating and over-evaluate cost coefficient, w of a jth wind energy turbine set _{av, j, t}for the actual capabilities of a jth wind energy turbine set t are exerted oneself, f _j(w _{av, j, t}) be the jth wind energy turbine set probability density function that actual capabilities are exerted oneself under corresponding wind-powered electricity generation prediction level, expression-form is the probability density function of the general distribution such as formula (1), and according to the corresponding general profile parameter of t, beta, gamma is determined; w _r,jfor the installed capacity of wind-driven power of a jth wind energy turbine set.

2.1.2 constraints

For the safe and stable operation of safeguards system, system should meet following constraints:

$\underset{i=1}{\overset{I}{\&Sigma;}}{p}_{i,t}+\underset{j=1}{\overset{J}{\&Sigma;}}{w}_{j,t}=L_t,\&ForAll;t---\left(11\right)$

${\mathrm{\&eta;}}_{i,t}{p}_{min,i}\&le;p_{i,t}\&le;{\mathrm{\&eta;}}_{i,t}{p}_{max,i},\&ForAll;i,t---\left(12\right)$

$0\&le;w_{j,t}\&le;w_{r,j},\&ForAll;j,t---\left(13\right)$

$p_{i,t}-p_{i,t-1}\&le;{\mathrm{\&eta;}}_{i,t-1}\&CenterDot;r_{u,max,i}+({\mathrm{\&eta;}}_{i,t}-{\mathrm{\&eta;}}_{i,t-1})p_{min,i}+(1-{\mathrm{\&eta;}}_{i,t})p_{max,i},\&ForAll;i,t---\left(14\right)$

$p_{i,t-1}-p_{i,t}\&le;{\mathrm{\&eta;}}_{i,t}\&CenterDot;r_{d,max,i}+({\mathrm{\&eta;}}_{i,t-1}-{\mathrm{\&eta;}}_{i,t})p_{min,i}+(1-{\mathrm{\&eta;}}_{i,t-1})p_{max,i},\&ForAll;i,t---\left(15\right)$

$0\&le;r_{u,i,t}\&le;min\{{\mathrm{\&eta;}}_{i,t}{p}_{max,i}-p_{i,t},{\mathrm{\&eta;}}_{i,t}{r}_{u,max,i}\},\&ForAll;i,t---\left(16\right)$

$0\&le;r_{d,i,t}\&le;min\{p_{i,t}-{\mathrm{\&eta;}}_{i,t}{p}_{min,i},{\mathrm{\&eta;}}_{i,t}{r}_{d,max,i}\},\&ForAll;i,t---\left(17\right)$

$\mathrm{Pr}\{\underset{i=1}{\overset{I}{\&Sigma;}}{r}_{u,i,t}\&GreaterEqual;\underset{j=1}{\overset{J}{\&Sigma;}}(w_{j,t}-w_{av,j,t})\}\&GreaterEqual;c_u,\&ForAll;t---\left(18\right)$

$\mathrm{Pr}\{\underset{i=1}{\overset{I}{\&Sigma;}}{r}_{d,i,t}\&GreaterEqual;\underset{j=1}{\overset{J}{\&Sigma;}}(w_{av,j,t}-w_{j,t})\}\&GreaterEqual;c_d,\&ForAll;t---\left(19\right)$

$-(1-\mathrm{\&mu;})\&CenterDot;F^{max}\&le;F_t\&le;(1-\mathrm{\&mu;})\&CenterDot;F^{max},\&ForAll;t---\left(20\right)$

Wherein, the power-balance that constraint (11) is system retrains, constraint (12), (13) are respectively the bound constraint of exerting oneself of fired power generating unit and wind energy turbine set, constraint (14), (15) are respectively the upwards climbing of fired power generating unit and downward Climing constant, the reserve capacity that constraint (16) ~ (19) are system retrains, and the trend that constraint (20) is system line retrains.L _tfor the total load of t system, η _i,tbe the on-off state of i-th fired power generating unit t: 1 represents that fired power generating unit is open state, and 0 represents that fired power generating unit is off-mode.W _r,jfor the installed capacity of a jth wind energy turbine set.R _{u, max, i}and r _{d, max, i}be respectively i-th fired power generating unit maximum creep speed up and down, p _{min, i}and p _{max, i}be minimum load and the maximum output of i-th fired power generating unit, r _{u, i, t}and r _{d, i, t}be the reserve capacity up and down of i-th fired power generating unit t, c _uand c _dbe respectively the confidence level that corresponding constraints meets.F _tfor the vector of each Line Flow of t, F ^maxfor the vector of each circuit maximum transfer capacity, the reserved transmission capacity of μ to be transmission line be wind-powered electricity generation fluctuation accounts for the ratio of each branch road maximum transfer capacity, and trend constraint DC flow model represents.

The conversion of 2.2 models and analysis

For above-mentioned stochastic and dynamic economic dispatch model, its decision variable is that the plan of fired power generating unit is exerted oneself and the plan of wind energy turbine set is exerted oneself, and stochastic variable is that the actual capabilities of wind energy turbine set are exerted oneself.Owing to containing stochastic variable in its target function and Reserve Constraint condition, cannot solve by direct classical algorithm.Therefore this trifle is based on the CDF of general distribution and contrafunctional closed analytical expression thereof, by correlation analysis and conversion, makes the stochastic and dynamic economic dispatch model based on general distribution be convenient to solve.

For the Reserve Constraint condition containing chance constraint, according to the contrafunctional analytical expression of general distribution CDF, formula (18), (19) can be converted into

$\underset{j=1}{\overset{J}{\&Sigma;}}{w}_{j,t}-\underset{i=1}{\overset{I}{\&Sigma;}}{r}_{u,i,t}\&le;F_{\mathrm{\&Sigma;},t}^{-1}(1-c_u),\&ForAll;t---\left(21\right)$

$-\underset{j=1}{\overset{J}{\&Sigma;}}{w}_{j,t}-\underset{i=1}{\overset{I}{\&Sigma;}}{r}_{d,i,t}\&le;-F_{\mathrm{\&Sigma;},t}^{-1}\left(c_d\right),\&ForAll;t---\left(22\right)$

Wherein, for the inverse function of the CDF that all wind energy turbine set actual capabilities of t are exerted oneself, described CDF is the cumulative distribution function of the general distribution such as formula (3), and according to the corresponding general profile parameter of t, beta, gamma is determined.Therefore, the constraints of Stochastic Optimization Model is all converted into linear restriction.

In target function, thermoelectricity fuel cost is quadratic function, and the punishment cost of wind-powered electricity generation is integral function.To target function C _alllocal derviation is asked to obtain

$\begin{array}{c}\frac{\&part;C_{all}(p_{i,t},w_{j,t})}{\&part;w_{j,t}}\\ =d_j=k_{un,j}{\&Integral;}_{w_{j,t}}^{w_{r,j}}{f}_j\left(w_{av,j,t}\right){\mathrm{dw}}_{av,j,t}+k_{ov,j}{\&Integral;}_0^{w_{j,t}}{f}_j\left(w_{av,j,t}\right){\mathrm{dw}}_{av,j,t}\\ =d_j=k_{un,j}{F}_j\left(w_{r,j}\right)-k_{ov,j}{F}_j\left(0\right)+(k_{un,j}+k_{ov,j})F_j\left(w_{j,t}\right)\end{array}---\left(23\right)$

$\frac{\&part;C_{all}(p_{i,t},w_{j,t})}{\&part;p_{i,t}}=2a_i{p}_{i,t}+b_i---\left(24\right)$

$\frac{{\&part;}^2{C}_{all}(p_{i,t},w_{j,t})}{\&part;{w^2}_{j,t}}=(k_{un,j}+k_{ov,j})f_j\left(w_{j,t}\right)---\left(25\right)$

$\frac{{\&part;}^2{C}_{all}(p_{i,t},w_{j,t})}{\&part;{p^2}_{i,t}}=2a_i---\left(26\right)$

$\frac{{\&part;}^2{C}_{all}(p_{i,t},w_{j,t})}{\&part;p_{i,t}\&part;w_{j,t}}=\frac{{\&part;}^2{C}_{all}(p_{i,t},w_{j,t})}{\&part;w_{j,t}\&part;p_{i,t}}=0---\left(27\right)$

Wherein, F _j(.) is the cumulative distribution function of corresponding stochastic variable, f _j(w _j,t) be w _j,tprobability density function.Because the second order local derviation of target function is all more than or equal to 0, therefore target function is convex function.By above-mentioned conversion and analysis, it is linear convex optimization problem that the stochastic and dynamic Economic Dispatch Problem based on general distribution finally changes into constraints, and the conventional optimized algorithms such as interior point method can be utilized to solve.

Solving of 2.3 models

Interior point method has a wide range of applications solving in convex optimization problem.Based on conversion and the analysis of above-mentioned model, propose a kind of quadratic programming-interior point method unified algorithm and solve corresponding convex optimization problem, namely utilize the solution of Novel Algorithm as primary iteration point, try to achieve globally optimal solution by interior point method successive iteration.

The plan of false wind electric field is exerted oneself as the predicted value of wind power, then can set up the certainty dynamic economic dispatch model based on prediction wind power, now target function becomes

$min\underset{t=1}{\overset{T}{\&Sigma;}}\underset{i=1}{\overset{I}{\&Sigma;}}(a_i{p}_{i,t}^2+b_i{p}_{i,t}+c_i)+C_{wind}---\left(28\right)$

Wherein, C _windfor the total cost of wind-powered electricity generation, directly can be added by formula (8), (9) and (10) and obtain, be constant term.

If the constraints of correspondence is constant, the certainty dynamic economic dispatch model based on prediction wind power be then made up of formula (28), (11)-(17), (21)-(23) has the form of quadratic programming, and ripe Novel Algorithm can be adopted to solve.Because the certainty economic dispatch model based on prediction wind power is consistent with the constraints of the random economic dispatch model based on general distribution, the solution of then being tried to achieve by Novel Algorithm also meets the constraints of Stochastic Optimization Model, and namely the solution of Novel Algorithm can be used as the primary iteration point of interior point method.Based on primary iteration point, solving constraints by interior point method is linear convex optimization problem, and then can obtain the optimal solution of the stochastic and dynamic economic dispatch based on general distribution, and the plan exporting fired power generating unit and wind energy turbine set is exerted oneself.

According to above model, embodiment provide based on the economic dispatch of the dynamic random a few days ago method of general distribution concrete solution procedure as shown in Figure 2:

Step 1, inputs prediction data a few days ago, comprises system loading prediction data (L a few days ago _t) and wind power prediction data (w a few days ago _{j, fcst, t}), and thermal power unit operation parameter (η _i,t, p _{min, i}, p _{max, i}, r _{d, max, i}, r _{u, max, i}), system line parameter (F ^max, μ), historical statistical data (under different wind power prediction level the general profile parameter of actual wind power, beta, gamma).

During concrete enforcement, can utilize the general profile parameter predicted in advance and obtain, beta, gamma, the embodiment of the present invention was divided into T moment by one day, and each moment t has corresponding general profile parameter, beta, gamma.During concrete enforcement, those skilled in the art can predetermined time length voluntarily, such as, be set to 15 minutes, then T=96.96 groups of general profile parameter of input prediction, beta, gamma, according to the general profile parameter that each moment is corresponding, beta, gamma and wind power prediction value w _{j, fcst, t}, the fired power generating unit plan can obtaining the corresponding moment at subsequent step is exerted oneself and the plan of wind energy turbine set is exerted oneself.During concrete enforcement, occasion length can be set neatly, corresponding solving is carried out to the required plan moment.

Step 2, the plan of false wind electric field is exerted oneself w _j,tfor wind power prediction value w _{j, fcst, t}, solve by formula (28) with Novel Algorithm, (11-17), the quadratic programming problem that (20-22) is formed, each fired power generating unit obtained based on prediction wind power is exerted oneself p _i,t ⁽⁰⁾.

Step 3, by wind power prediction value w _{j, fcst, t}to exert oneself p with each fired power generating unit solved _i,t ⁽⁰⁾as the primary iteration point x of interior point method ⁽⁰⁾.

Step 4, arranges convergence criterion parameter ε and the maximum iteration time N of interior point method _iter; During concrete enforcement, those skilled in the art can preset convergence criterion parameter ε and maximum iteration time N voluntarily _itervalue, such as embodiment gets maximum iteration time N _iterbe 1000, convergence criterion parameter ε is 0.001.

Step 5, constraints after utilizing the nonlinear optimization solved function fmincon in MATLAB to solve conversion is linear convex optimization problem, namely by formula (6), (11-17), (20-22) the convex optimization problem formed, wherein algorithm selects interior point method (interior-point).First according to current primary iteration point x ⁽⁰⁾calculate variable (p _i,t, w _j,t, r _{u, i, t}and r _{d, i, t}), if do not meet iteration termination condition, then according to current variable (p _i,t, w _j,t, r _{u, i, t}and r _{d, i, t}) continue to solve and obtain new (p _i,t, w _j,t, r _{u, i, t}and r _{d, i, t}), until meet iteration termination condition.The maximum iteration time N of embodiment _iterbe 1000, as the changing value C of target function _allbe less than 0.001 or variable (p _i,t, w _j,t, r _{u, i, t}and r _{d, i, t}) the maximum of changing value iteration stopping when being less than 0.001, enter step 6.

Step 6, exports fired power generating unit plan and to exert oneself (p _i,t) and the plan of wind energy turbine set to exert oneself (w _j,t).According to variable (p during step 5 finishing iteration _i,t, w _j,t, r _{u, i, t}and r _{d, i, t}), final fired power generating unit plan can be obtained and to exert oneself and the plan of wind energy turbine set is exerted oneself.

During concrete enforcement, software engineering can be adopted to realize the automatic operation of above process, modular mode also can be adopted to provide corresponding system.The embodiment of the present invention provides the probability density function of cloth, and according to the corresponding general profile parameter of t, beta, gamma is determined.

The present invention is corresponding provides a kind of Electrical Power System Dynamic based on general distribution random economic dispatch system, comprises with lower module:

Input module, for inputting system loading prediction data and a few days ago wind power prediction data, thermal power unit operation parameter, system line parameter, historical statistical data a few days ago, described historical statistical data comprises the general profile parameter of actual wind power under different wind power prediction level, beta, gamma;

Tentatively solve module, to exert oneself w for the plan of false wind electric field _j,tfor wind power prediction value w _{j, fcst, t}, wind power prediction value w _{j, fcst, t}thered is provided by wind power prediction data a few days ago, solve by formula (28) with Novel Algorithm, (11-17), the quadratic programming problem that (20-22) is formed, obtain exerting oneself p based on each fired power generating unit of prediction wind power _i,t ⁽⁰⁾;

Initialization module, for by wind power prediction value w _{j, fcst, t}to exert oneself p with each fired power generating unit solved _i,t ⁽⁰⁾as the primary iteration point x of interior point method ⁽⁰⁾;

Condition setting module, for arranging convergence criterion parameter ε and the maximum iteration time N of interior point method _iter;

Iteration module, constraints after transforming for utilizing interior point method iterative is linear convex optimization problem, namely by formula (6), (11-17), (20-22) the convex optimization problem formed, until meet convergence criterion parameter ε or maximum iteration time N _itertime iteration stopping, order output module work

Output module, for the iteration result according to iteration module, the plan exporting fired power generating unit and wind energy turbine set is exerted oneself.

Each module realizes can see foregoing teachings, and it will not go into details in the present invention.

Finally, for illustrating for the purpose of the technology of the present invention effect, sample calculation analysis is provided:

3.1 optimum configurations

This section, for the IEEE30 node system containing 1 wind energy turbine set, verifies the validity of institute's extracting method herein.IEEE30 node is one of international standard test macro, herein be increase the IEEE30 node system after wind energy turbine set.As shown in Figure 3, wherein G1 ~ G6 is fired power generating unit to the network topological diagram of amended IEEE30 node system, and W1 is the 1st wind energy turbine set, and 1 ~ 30 is system node label.In system, installed capacity of wind-driven power is 100MW, and original wind-powered electricity generation is data from Ireland.Fired power generating unit parameter is as shown in table 1, and the minimum technology that wherein PGmin, PGmax are respectively fired power generating unit is exerted oneself and maximum technology is exerted oneself, and a, b, c are respectively the fuel cost coefficient of fired power generating unit.Line parameter circuit value is shown in document [ZhangS, SongY, HuZ, etal.Robustoptimizationmethodbasedonscenarioanalysisforu nitcommitmentconsideringwinduncertainties [C] .PowerandEnergySocietyGeneralMeeting, SanDiego, CA, USA, 2011.], wherein, the maximum transfer capacity of circuit 1-2 (Line1) and 9-10 (Line14) is respectively 110MW and 105MW, the maximum transfer capacity of other circuit is 100MW, all circuits are that the ratio μ that the reserved transmission capacity of wind-powered electricity generation fluctuation accounts for corresponding line maximum transfer capacity all gets 5%, the confidence level c of Reserve Constraint _u, c _dall get 95%.The cost coefficient of underestimating of wind-powered electricity generation gets 80 $/MWh, and the cost coefficient of over-evaluating of wind-powered electricity generation gets 120 $/MWh, ignores the basic operating cost of wind-powered electricity generation.For the system total load curve of stochastic and dynamic economic dispatch a few days ago and wind power prediction curve (15 minutes points) as shown in Figure 4, corresponding fired power generating unit switching on and shutting down plan is a few days ago as shown in table 2.

Table 1

Table 2

3.2 general distributions describe the validity of wind power

Statistical analysis is carried out to Ireland history wind power data of 2 years.First its installed capacity is equivalent to 100MW, then according to the predicted value of wind power, branch mailbox is carried out to history wind power data, and general distribution and normal distyribution function matching are utilized respectively to the distribution of the actual wind power in the pre-measuring tank of difference.Under corresponding different prediction level, the root-mean-square error of the general fitting of distribution parameter of actual wind power, normal distribution fitting parameter and correspondence thereof is as shown in table 3.

Table 3

As shown in Table 3, when predicting that wind power is less (the 1st, 2 pre-measuring tank), actual wind power output is by the restriction of wind energy turbine set minimum load (0MW), the parameter beta (227.8 of general distribution, 349.7) much larger than 1, actual distribution presents obvious right avertence state, compared to normal distribution, the root-mean-square error of general distribution is less, fitting effect is better, corresponding fitting effect contrast as shown in fig. 5-1 (for the 2nd pre-measuring tank), which provides actual distribution histogram, general fitting of distribution curve, normal distribution matched curve; When predicting that wind power is placed in the middle (for the 10th pre-measuring tank), the parameter beta of general distribution is close to 1, the distribution of actual wind power distributes close to without inclined, the root-mean-square error of normal distribution and general distribution is all less, both fitting effect gaps are little, and corresponding fitting effect is to such as shown in Fig. 5-2; When predicting that wind power is larger (the 19th, 20 pre-measuring tank), actual wind power output is by the restriction of wind energy turbine set maximum output (installed capacity 100MW), the parameter beta (0.1836 of general distribution, 0.2743) 1 is significantly less than, actual distribution presents obvious left avertence state, and compared to normal distribution, the root-mean-square error of general distribution is less, fitting effect is better, and corresponding fitting effect is to such as shown in Fig. 5-3 (for the 19th pre-measuring tank).

As the above analysis, compared to normal distribution, general distribution can according to the distribution character of actual wind power, by the adjustment of parameter beta can the prediction of matching preferably wind power less or larger time, actual wind power distributes the offset characteristic had.Therefore, the uncertainty of wind power can be considered more accurately based on the stochastic and dynamic economic dispatch of general distribution.

3.3 random economic dispatch interpretations of result

This trifle, by compared with the stochastic and dynamic economic dispatch method based on normal distribution, demonstrates the validity of the stochastic and dynamic economic dispatch method based on general distribution.

Based on wind-powered electricity generation dispatch curve corresponding to the random economic dispatch method of normal distribution and general distribution as shown in Figure 6, prediction and actual wind-powered electricity generation curve, wind-powered electricity generation dispatch curve based on normal distribution and general distribution correspondence is which provided.Two kinds of methods all can (in the upper and lower limit of wind power) be optimized it and exert oneself in 90% confidential interval of actual wind power fluctuation range.But compared to normal distribution, general distribution can the distribution of the actual wind power of matching better, thus can consider the uncertainty of wind power more accurately based on the stochastic and dynamic economic dispatch model of general distribution, the scheduling result of its correspondence is also more effective.

3.3.1 system reserve capacity analysis

The system reserve capacity that two kinds of methods are corresponding and wind power fluctuate to the demand of system reserve respectively as shown in Figure 7, Figure 8, which provide wind-powered electricity generation fluctuation to demand upwards for subsequent use with to demand for subsequent use downwards.

As seen from the figure, the downward reserve capacity abundance (being 701.52MW) that the system of two kinds of methods is total, is enough to the upwards fluctuation tackling wind power.But because normal distribution and general distribution all exist error of fitting, when upwards reserve capacity is all inadequate for system, two kinds of methods are all not enough to the downward fluctuation tackling wind power completely.

The reserved upwards reserve capacity of two kinds of methods is as shown in table 4.Reserved total upwards reserve capacity fluctuate downwards a little less than the method based on general distribution although the method based on normal distribution is wind-powered electricity generation, but its upwards total vacancy for subsequent use be 38.96MW, obviously be greater than total vacancy 16.35MW of the method based on general distribution, and the maximum vacancy of its correspondence is 6.80MW (the t=8h moment), be also obviously greater than the maximum vacancy 1.52MW (t=10h moment) of the method based on general distribution.When actual wind power output be less than plan exert oneself time, operation plan corresponding to the method based on normal distribution may make system be difficult to reply wind-powered electricity generation due to upwards deficiency for subsequent use significantly to fluctuate downwards.

Table 4

Because general distribution can the distribution situation of actual wind power under the different wind-powered electricity generation prediction level of matching more exactly, the error of fitting of its correspondence is less than normal distribution, thus the economic dispatch model based on general distribution can more reasonably for the reserved reserve capacity of wind-powered electricity generation fluctuation be to tackle the uncertainty of wind-powered electricity generation, be convenient to the adjustment of in a few days operation plan, run with safeguards system economic security, avoid system to abandon wind, cutting load as much as possible.

3.3.2 cost analysis

Cost corresponding to two kinds of methods is as shown in table 5.As shown in Table 5, thermoelectricity fuel cost corresponding to the method based on normal distribution is lower, this is because it is reserved upwards for subsequent use few, but the method does not describe the distribution of wind power exactly, so the wind-powered electricity generation punishment cost of its correspondence is higher.In general, based on the total cost of the method for general distribution lower than the method based on normal distribution, thus there is better economy.Consider economy and the fail safe of system, the stochastic and dynamic economic dispatch method based on general distribution can provide more effective reference for system coordinator.

Table 5

For the random economic dispatch method based on general distribution, when the confidence level of system reserve constraint satisfaction is different, each cost change of system as shown in Figure 9, comprises total cost, thermoelectricity fuel cost, the change of wind-powered electricity generation punishment cost.Along with the raising of confidence level, fired power generating unit need adjust its optimal output with reserved enough reserve capacitys to meet Reserve Constraint condition, and this will increase the fuel cost of fired power generating unit.And along with the increase of system reserve capacity, wind energy turbine set can be optimized its plan further and exert oneself, so reduce wind-powered electricity generation due to low, over-evaluate the average punishment cost brought.But under same fired power generating unit switching on and shutting down plan, whole system is reduce risk to increase reserve capacity with the safe and stable operation of safeguards system, and this also result in the increase of total cost.

3.3.3 trend is about beam analysis

From the result of Load flow calculation, circuit 1-2 (Line1) is owing to connecting fired power generating unit G1 and G2, circuit 9-10 (Line14) owing to connecting G5 and heavy loading district, and the risk that corresponding Line Flow is out-of-limit is larger.According to the distribution character of not wind power in the same time, the wind power curve that stochastic simulation 10000 may occur, obtain the probability that Line Flow meets corresponding constraints, the Comparative result of two kinds of methods is as shown in table 6.As seen from the figure, method based on general distribution and the method based on normal distribution all can meet trend constraints with larger probability, avoid system load flow out-of-limit, and the probability of institute's extracting method trend constraint satisfaction is all greater than the dispatching method based on normal distribution, thus demonstrate the validity of institute's extracting method herein further herein.

Table 6

The present invention is on the basis analyzing the distribution of actual wind power, establish the economic dispatch of the dynamic random a few days ago model that consideration wind-powered electricity generation based on general distribution is low, over-evaluate cost, propose a kind of quadratic programming-interior point method unified algorithm and solve corresponding dynamic random Economic Dispatch Problem.Based on IEEE30 bus test system, carried out simulating, verifying, result shows:

1), compared with describing the method for wind power with conventional normal distribution, the distribution of actual wind power under different wind power prediction value can be described more accurately based on the method for general distribution.

2) compared to the dynamic random economic dispatch method based on normal distribution, dynamic random economic dispatch method based on general distribution can consider the uncertainty of wind power more accurately, and then suitable reserve capacity can be reserved for the fluctuation of wind power, ensure that the trend of circuit meets constraints, to fluctuate the impact brought guaranteeing to reduce as much as possible under the prerequisite that system is comparatively safe wind-powered electricity generation, thus reduce the total operating cost of system.

It is emphasized that; embodiment of the present invention is illustrative; instead of it is determinate; therefore the present invention is not limited to the embodiment described in embodiment; every other execution modes drawn by those skilled in the art's technical scheme according to the present invention, belong to the scope of protection of the invention equally.

Claims (10)

1., based on the random economic dispatch method of Electrical Power System Dynamic of general distribution, it is characterized in that, comprise the following steps:

Step 1, input system loading prediction data and a few days ago wind power prediction data, thermal power unit operation parameter, system line parameter, historical statistical data a few days ago, described historical statistical data comprises the general profile parameter of actual wind power under different wind power prediction level, beta, gamma;

Step 2, if p _i,tbe exerting oneself of i-th fired power generating unit t, fired power generating unit add up to I, i=1,2 ..., I, w _j,tfor the plan of a jth wind energy turbine set t is exerted oneself, wind energy turbine set add up to J, j=1,2 ..., J, T are the sum in moment, t=1,2 ..., T,

The plan of false wind electric field is exerted oneself w _j,tfor wind power prediction value w _{j, fcst, t}, wind power prediction value w _{j, fcst, t}thered is provided by wind power prediction data a few days ago, solve following quadratic programming problem with Novel Algorithm, obtain exerting oneself p based on each fired power generating unit of prediction wind power _i,t ⁽⁰⁾,

$min\underset{t=1}{\overset{T}{\&Sigma;}}\underset{i=1}{\overset{I}{\&Sigma;}}(a_i{p}_{i,t}^2+b_i{p}_{i,t}+c_i)+C_{wind}$ (formula one)

Wherein, C _windfor the total cost of wind-powered electricity generation, a _i, b _i, c _iit is the fuel cost coefficient of i-th fired power generating unit;

$\underset{i=1}{\overset{I}{\&Sigma;}}{p}_{i,t}+\underset{j=1}{\overset{J}{\&Sigma;}}{w}_{j,t}=L_t,\&ForAll;t$ (formula two)

${\mathrm{\&eta;}}_{i,t}{p}_{min,i}\&le;p_{i,t}\&le;{\mathrm{\&eta;}}_{i,t}{p}_{max,i},\&ForAll;i,t$ (formula three)

$0\&le;w_{j,t}\&le;w_{r,j},\&ForAll;j,t$ (formula four)

$p_{i,t}-p_{i,t-1}\&le;{\mathrm{\&eta;}}_{i,t-1}\&CenterDot;r_{u,max,i}+({\mathrm{\&eta;}}_{i,t}-{\mathrm{\&eta;}}_{i,t-1})p_{min,i}+(1-{\mathrm{\&eta;}}_{i,t})p_{max,i},\&ForAll;i,t$ (formula five)

$p_{i,t-1}-p_{i,t}\&le;{\mathrm{\&eta;}}_{i,t}\&CenterDot;r_{d,max,i}+({\mathrm{\&eta;}}_{i,t-1}-{\mathrm{\&eta;}}_{i,t})p_{min,i}+(1-{\mathrm{\&eta;}}_{i,t-1})p_{max,i},\&ForAll;i,t$ (formula six)

$0\&le;r_{u,i,t}\&le;min\{{\mathrm{\&eta;}}_{i,t}{p}_{max,i}-p_{i,t},{\mathrm{\&eta;}}_{i,t}{r}_{u,max,i}\},\&ForAll;i,t$ (formula seven)

$0\&le;r_{d,i,t}\&le;min\{p_{i,t}-{\mathrm{\&eta;}}_{i,t}{p}_{min,i},{\mathrm{\&eta;}}_{i,t}{r}_{d,max,i}\},\&ForAll;i,t$ (formula eight)

Wherein, L _tfor the total load of t system, provided by system loading prediction data a few days ago; η _i,tbe the on-off state of i-th fired power generating unit t, r _{u, max, i}and r _{d, max, i}be respectively i-th fired power generating unit maximum creep speed up and down, p _{min, i}and p _{max, i}be minimum load and the maximum output of i-th fired power generating unit, provided by thermal power unit operation parameter; w _r,jfor the installed capacity of wind-driven power of a jth wind energy turbine set, r _{u, i, t}and r _{d, i, t}it is the reserve capacity up and down of i-th fired power generating unit t;

$-(1-\mathrm{\&mu;})\&CenterDot;F^{max}\&le;F_t\&le;(1-\mathrm{\&mu;})\&CenterDot;F^{max},\&ForAll;t$ (formula nine)

Wherein, F _tfor the vector of each Line Flow of t; F ^maxfor the vector of each circuit maximum transfer capacity, the reserved transmission capacity of μ to be transmission line be wind-powered electricity generation fluctuation accounts for the ratio of each branch road maximum transfer capacity, is provided by system line parameter;

$\underset{j=1}{\overset{J}{\&Sigma;}}{w}_{j,t}-\underset{i=1}{\overset{I}{\&Sigma;}}{r}_{u,i,t}\&le;F_{\mathrm{\&Sigma;},t}^{-1}(1-c_u),\&ForAll;t$ (formula ten)

$-\underset{j=1}{\overset{J}{\&Sigma;}}{w}_{j,t}-\underset{i=1}{\overset{I}{\&Sigma;}}{r}_{d,i,t}\&le;-F_{\mathrm{\&Sigma;},t}^{-1}\left(c_d\right),\&ForAll;t$ (formula 11)

Wherein, for the inverse function of the CDF that all wind energy turbine set actual capabilities of t are exerted oneself, c _uand c _dbe respectively the confidence level that corresponding constraints meets; Described CDF is the cumulative distribution function of general distribution, and according to the corresponding general profile parameter of t, beta, gamma is determined;

Step 3, by wind power prediction value w _{j, fcst, t}to exert oneself p with each fired power generating unit solved _i,t ⁽⁰⁾as the primary iteration point x of interior point method ⁽⁰⁾;

Step 4, arranges convergence criterion parameter ε and the maximum iteration time N of interior point method _iter;

Step 5, the constraints after utilizing interior point method iterative to transform is linear convex optimization problem, and namely by formula 12, the convex optimization problem that formula two-Shi 11 is formed, until meet convergence criterion parameter ε or maximum iteration time N _itertime iteration stopping, enter step 6,

${\mathrm{minC}}_{all}(p_{i,t},w_{j,t})=\underset{t=1}{\overset{T}{\&Sigma;}}\underset{i=1}{\overset{I}{\&Sigma;}}{C}_{g,i,t}\left(p_{i,t}\right)+\underset{t=1}{\overset{T}{\&Sigma;}}\underset{j=1}{\overset{J}{\&Sigma;}}\left(C_{w,j,t}\right(w_{j,t})+C_{un,j,t}(w_{j,t})+C_{ov,j,t}(w_{j,t}\left)\right)$ (formula 12)

Wherein, C _allfor the operating cost that system is total, C _{g, i, t}for the fuel cost of t i-th fired power generating unit, C _{w, j, t}for the operating cost of a t jth wind energy turbine set, C _{un, j, t}on average cost is underestimated, C for a t jth wind energy turbine set wind power prediction _{ov, j, t}on average cost is over-evaluated for a t jth wind energy turbine set wind power prediction;

Step 6, according to the iteration result of step 5, the plan exporting fired power generating unit and wind energy turbine set is exerted oneself.

2., according to claim 1 based on the random economic dispatch method of Electrical Power System Dynamic of general distribution, it is characterized in that: described fuel cost C _{g, i, t}calculate in the following ways,

$C_{g,i,t}\left(p_{i,t}\right)=a_i{p}_{i,t}^2+b_i{p}_{i,t}+c_i$ (formula 13)

Wherein, p _i,tbe exerting oneself of i-th fired power generating unit t, a _i, b _i, c _iit is the fuel cost coefficient of i-th fired power generating unit.

3., according to claim 1 based on the random economic dispatch method of Electrical Power System Dynamic of general distribution, it is characterized in that: described operating cost C _{w, j, t}calculate in the following ways,

C _{w, j, t}(w _j,t)=d _jw _j,t(formula 14)

Wherein, w _j,tfor the plan of a jth wind energy turbine set t is exerted oneself, d _jfor the operating cost coefficient of a jth wind energy turbine set.

4. according to claim 1 based on the random economic dispatch method of Electrical Power System Dynamic of general distribution, it is characterized in that: describedly on average underestimate cost C _{un, j, t}what adopt wind energy turbine set on average abandons eolian, calculates in the following ways,

$C_{un,j,t}\left(w_{j,t}\right)=k_{un,j}\underset{w_{j,t}}{\overset{w_{r,j}}{\&Integral;}}(w_{av,j,t}-w_{j,t})\&CenterDot;f_j\left(w_{av,j,t}\right){\mathrm{dw}}_{av,j,t}$ (formula 15)

Wherein, k _{un, j}for a jth wind energy turbine set corresponding underestimate cost coefficient, w _{av, j, t}for the actual capabilities of a jth wind energy turbine set t are exerted oneself, f _j(w _{av, j, t}) be the jth wind energy turbine set probability density function that actual capabilities are exerted oneself under corresponding wind-powered electricity generation prediction level, expression-form is the probability density function of general distribution, and according to the corresponding general profile parameter of t, beta, gamma is determined.

5. according to claim 1 based on the random economic dispatch method of Electrical Power System Dynamic of general distribution, it is characterized in that: describedly on average over-evaluate cost C _{ov, j, t}the average stand-by cost of employing system, calculates in the following ways,

$C_{ov,j,t}\left(w_{j,i}\right)=k_{ov,j}\underset{0}{\overset{w_{j,t}}{\&Integral;}}(w_{j,t}-w_{av,j,t})\&CenterDot;f_j\left(w_{av,j,t}\right){\mathrm{dw}}_{av,j,t}$ (formula 16)

Wherein, k _{ov, j}for a jth wind energy turbine set corresponding over-evaluate cost coefficient, w _{av, j, t}for the actual capabilities of a jth wind energy turbine set t are exerted oneself, f _j(w _{av, j, t}) be the jth wind energy turbine set probability density function that actual capabilities are exerted oneself under corresponding wind-powered electricity generation prediction level, expression-form is the probability density function of general distribution, and according to the corresponding general profile parameter of t, beta, gamma is determined.

6., based on the random economic dispatch system of Electrical Power System Dynamic of general distribution, it is characterized in that, comprise with lower module:

Input module, for inputting system loading prediction data and a few days ago wind power prediction data, thermal power unit operation parameter, system line parameter, historical statistical data a few days ago, described historical statistical data comprises the general profile parameter of actual wind power under different wind power prediction level, beta, gamma;

Tentatively solve module, for establishing p _i,tbe exerting oneself of i-th fired power generating unit t, fired power generating unit add up to I, i=1,2 ..., I, w _j,tfor the plan of a jth wind energy turbine set t is exerted oneself, wind energy turbine set add up to J, j=1,2 ..., J, T are the sum in moment, t=1,2 ..., T,

The plan of false wind electric field is exerted oneself w _j,tfor wind power prediction value w _{j, fcst, t}, wind power prediction value w _{j, fcst, t}thered is provided by wind power prediction data a few days ago, solve following quadratic programming problem with Novel Algorithm, obtain exerting oneself p based on each fired power generating unit of prediction wind power _i,t ⁽⁰⁾,

$min\underset{t=1}{\overset{T}{\&Sigma;}}\underset{i=1}{\overset{I}{\&Sigma;}}(a_i{p}_{i,t}^2+b_i{p}_{i,t}+c_i)+C_{wind}$ (formula one)

Wherein, C _windfor the total cost of wind-powered electricity generation, a _i, b _i, c _iit is the fuel cost coefficient of i-th fired power generating unit;

$\underset{i=1}{\overset{I}{\&Sigma;}}{p}_{i,t}+\underset{j=1}{\overset{J}{\&Sigma;}}{w}_{j,t}=L_t,\&ForAll;t$ (formula two)

${\mathrm{\&eta;}}_{i,t}{p}_{min,i}\&le;p_{i,t}\&le;{\mathrm{\&eta;}}_{i,t}{p}_{max,i},\&ForAll;i,t$ (formula three)

$0\&le;w_{j,t}\&le;w_{r,j},\&ForAll;j,t$ (formula four)

$p_{i,t}-p_{i,t-1}\&le;{\mathrm{\&eta;}}_{i,t-1}\&CenterDot;r_{u,max,i}+({\mathrm{\&eta;}}_{i,t}-{\mathrm{\&eta;}}_{i,t-1})p_{min,i}+(1-{\mathrm{\&eta;}}_{i,t})p_{max,i},\&ForAll;i,t$ (formula five)

$p_{i,t-1}-p_{i,t}\&le;{\mathrm{\&eta;}}_{i,t}\&CenterDot;r_{d,max,i}+({\mathrm{\&eta;}}_{i,t-1}-{\mathrm{\&eta;}}_{i,t})p_{min,i}+(1-{\mathrm{\&eta;}}_{i,t-1})p_{max,i},\&ForAll;i,t$ (formula six)

$0\&le;r_{u,i,t}\&le;min\{{\mathrm{\&eta;}}_{i,t}{p}_{max,i}-p_{i,t},{\mathrm{\&eta;}}_{i,t}{r}_{u,max,i}\},\&ForAll;i,t$ (formula seven)

$0\&le;r_{d,i,t}\&le;min\{p_{i,t}-{\mathrm{\&eta;}}_{i,t}{p}_{min,i},{\mathrm{\&eta;}}_{i,t}{r}_{d,max,i}\},\&ForAll;i,t$ (formula eight)

Wherein, L _tfor the total load of t system, provided by system loading prediction data a few days ago; η _i,tbe the on-off state of i-th fired power generating unit t, r _{u, max, i}and r _{d, max, i}be respectively i-th fired power generating unit maximum creep speed up and down, p _{min, i}and p _{max, i}be minimum load and the maximum output of i-th fired power generating unit, provided by thermal power unit operation parameter; w _r,jfor the installed capacity of wind-driven power of a jth wind energy turbine set, r _{u, i, t}and r _{d, i, t}it is the reserve capacity up and down of i-th fired power generating unit t;

$-(1-\mathrm{\&mu;})\&CenterDot;F^{max}\&le;F_t\&le;(1-\mathrm{\&mu;})\&CenterDot;F^{max},\&ForAll;t$ (formula nine)

Wherein, F _tfor the vector of each Line Flow of t; F ^maxfor the vector of each circuit maximum transfer capacity, the reserved transmission capacity of μ to be transmission line be wind-powered electricity generation fluctuation accounts for the ratio of each branch road maximum transfer capacity, is provided by system line parameter;

$\underset{j=1}{\overset{J}{\&Sigma;}}{w}_{j,t}-\underset{i=1}{\overset{I}{\&Sigma;}}{r}_{u,i,t}\&le;F_{\mathrm{\&Sigma;},t}^{-1}(1-c_u),\&ForAll;t$ (formula ten)

$-\underset{j=1}{\overset{J}{\&Sigma;}}{w}_{j,t}-\underset{i=1}{\overset{I}{\&Sigma;}}{r}_{d,i,t}\&le;-F_{\mathrm{\&Sigma;},t}^{-1}\left(c_d\right),\&ForAll;t$ (formula 11)

Wherein, for the inverse function of the CDF that all wind energy turbine set actual capabilities of t are exerted oneself, c _uand c _dbe respectively the confidence level that corresponding constraints meets; Described CDF is the cumulative distribution function of general distribution, and according to the corresponding general profile parameter of t, beta, gamma is determined;

Initialization module, for by wind power prediction value w _{j, fcst, t}to exert oneself p with each fired power generating unit solved _i,t ⁽⁰⁾as the primary iteration point x of interior point method ⁽⁰⁾;

Condition setting module, for arranging convergence criterion parameter ε and the maximum iteration time N of interior point method _iter;

Iteration module, the constraints after transforming for utilizing interior point method iterative is linear convex optimization problem, and namely by formula 12, the convex optimization problem that formula two-Shi 11 is formed, until meet convergence criterion parameter ε or maximum iteration time N _itertime iteration stopping, order output module work,

${\mathrm{minC}}_{all}(p_{i,t},w_{j,t})=\underset{t=1}{\overset{T}{\&Sigma;}}\underset{i=1}{\overset{I}{\&Sigma;}}{C}_{g,i,t}\left(p_{i,t}\right)+\underset{t=1}{\overset{T}{\&Sigma;}}\underset{j=1}{\overset{J}{\&Sigma;}}\left(C_{w,j,t}\right(w_{j,t})+C_{un,j,t}(w_{j,t})+C_{ov,j,t}(w_{j,t}\left)\right)$ (formula 12)

Wherein, C _allfor the operating cost that system is total, C _{g, i, t}for the fuel cost of t i-th fired power generating unit, C _{w, j, t}for the operating cost of a t jth wind energy turbine set, C _{un, j, t}on average cost is underestimated, C for a t jth wind energy turbine set wind power prediction _{ov, j, t}on average cost is over-evaluated for a t jth wind energy turbine set wind power prediction;

Output module, for the iteration result according to iteration module, the plan exporting fired power generating unit and wind energy turbine set is exerted oneself.

7., according to claim 6 based on the random economic dispatch system of Electrical Power System Dynamic of general distribution, it is characterized in that: described fuel cost C _{g, i, t}calculate in the following ways,

$C_{g,i,t}\left(p_{i,t}\right)=a_i{p}_{i,t}^2+b_i{p}_{i,t}+c_i$ (formula 13)

Wherein, p _i,tbe exerting oneself of i-th fired power generating unit t, a _i, b _i, c _iit is the fuel cost coefficient of i-th fired power generating unit.

8., according to claim 6 based on the random economic dispatch system of Electrical Power System Dynamic of general distribution, it is characterized in that: described operating cost C _{w, j, t}calculate in the following ways,

C _{w, j, t}(w _j,t)=d _jw _j,t(formula 14)

Wherein, w _j,tfor the plan of a jth wind energy turbine set t is exerted oneself, d _jfor the operating cost coefficient of a jth wind energy turbine set.

9. according to claim 6 based on the random economic dispatch system of Electrical Power System Dynamic of general distribution, it is characterized in that: describedly on average underestimate cost C _{un, j, t}what adopt wind energy turbine set on average abandons eolian, calculates in the following ways,

$C_{un,j,t}\left(w_{j,t}\right)=k_{un,j}\underset{w_{j,t}}{\overset{w_{r,j}}{\&Integral;}}(w_{av,j,t}-w_{j,t})\&CenterDot;f_j\left(w_{av,j,t}\right){\mathrm{dw}}_{av,j,t}$ (formula 15)

Wherein, k _{un, j}for a jth wind energy turbine set corresponding underestimate cost coefficient, w _{av, j, t}for the actual capabilities of a jth wind energy turbine set t are exerted oneself, f _j(w _{av, j, t}) be the jth wind energy turbine set probability density function that actual capabilities are exerted oneself under corresponding wind-powered electricity generation prediction level, expression-form is the probability density function of general distribution, and according to the corresponding general profile parameter of t, beta, gamma is determined.

10. according to claim 6 based on the random economic dispatch system of Electrical Power System Dynamic of general distribution, it is characterized in that: describedly on average over-evaluate cost C _{ov, j, t}the average stand-by cost of employing system, calculates in the following ways,

$C_{ov,j,t}\left(w_{j,i}\right)=k_{ov,j}\underset{0}{\overset{w_{j,t}}{\&Integral;}}(w_{j,t}-w_{av,j,t})\&CenterDot;f_j\left(w_{av,j,t}\right){\mathrm{dw}}_{av,j,t}$ (formula 16)

Wherein, k _{ov, j}for a jth wind energy turbine set corresponding over-evaluate cost coefficient, w _{av, j, t}for the actual capabilities of a jth wind energy turbine set t are exerted oneself, f _j(w _{av, j, t}) be the jth wind energy turbine set probability density function that actual capabilities are exerted oneself under corresponding wind-powered electricity generation prediction level, expression-form is the probability density function of general distribution, and according to the corresponding general profile parameter of t, beta, gamma is determined.